I wish we could somehow force people to use a standard toolkit again, but I think the cat is out of the bag.
I remember back in the 90's, when Apple and Microsoft bragged about having big UI labs where they tried to take a scientific approach to UI design where they'd try out minor UI tweaks on users and observe what worked and what didn't.
Looking at the designs coming out of both companies now (but mostly Apple) it's extremely clear they they are designed just to look "fresh" and "pretty" with no real justification behind it.
> I wish we could somehow force people to use a standard toolkit again
To be fair, in the 90's, there were plenty of crazily themed programs. Basically any media playing app seemed required to have some insane theme (WinAmp is fondly remembered but that was also a non-standard UI, which was so poorly-designed they needed a button to double the size of it so you could hit some buttons easier). Every PC sound card came with a suite of programs that tried to mimic the appearance of a home Hi-Fi system complete with 7-segment displays and fake cassette tape reels.
And yes, on text you'd want a minimal UI. You don't see books with flashy things everywhere competing for your attention, maybe excepting kids' books.
My guess is that, as someone already pointed about design doing a full circle again in a few years, it already has happened in the web several times - plain text and 'brutalist' design, then the native UI controls that had to be used within web interfaces, then skeumorphic design to compensate for the native UI controls, then you could style those UI controls as you wished, then flattening all the design again because most of web is text.
Bingo. This is exactly it for me, right here. Being able to tell what I can click on and expect it to do something, at a glance and without having to guess is the killer feature, IMO. How did we lose our way so badly that "can I click on this thing?" even became a reasonable question to ask of a modern UI?
You can scan an interface of this fashion and be almost certain about what is a button, what is a scroll bar, what is a text input, etcetera.
While with Windows 8+ the flat UI lets a huge space for ambiguity and causes a bigger cognitive load to the user, who is forced to tell things between flat rectangles and lines.